Re: Gold plating the entire PCB

Horrible little buggers. Insertion force is insane. This will be populated with SMBs, since it'll only go to a couple of GHz.

MCX's are nice, and fit into our standard SMA/SMB/LEMO footprint. Cables are cheap on ebay. We just have a lot of SMB stuff around here.

This was done by Cirexx, one of our "regular" board houses. It's FR4. The test trace impedances came out to be 46.2 ohms on layer 1, 57 on layer 3, and 47.2 on layer 4 (L2 is ground plane). Mediocre shooting. I added some right angles and beveled corners for fun, but they're totally invisible on the 20 GHz TDR.

Electronic Interconnect does cheap 4 layer protos with 14/28/14 mil dielectrics. I could have planned for that and saved some money. Now I know.

We only need 1 GHz for now, to calibrate some photodiode gain paths. That should be easy... I got pretty good 1 GHz results on an x-acto hacked board. I'm expecting usable 2 GHz or so here. Heck, a trimpot is good to 500 MHz.

John

Cool, please let us know what the measurement is; I'll honestly be glad to be wrong in my estimates here. :-)

Anything designed in the UK between 1960 and 1975 leaks oil.

Cheers,

Phil "Former TR-7 owner" Hobbs

I planned it so it can be band-sawed into a bunch of little tiles.

The two CPW thingies near the top will use edge-launch connectors, SMA or SMB. The thru-hole connectors will be SMBs, because they're easy to connect and disconnect.

John

I was always an MG guy myself, but I know what you mean. I have the GEC schematics, and they're insane.

Soichiro Honda should have got the Nobel Prize in physics for his revelation: let's split the motorcycle crankcase *horizontally*

John

Hmm, Dimmie, this *is* sci.ELECTRONICS.design. You don't think a discussion of metastability belongs in an ELECTRONICS discussion?

Hmm, that's a new one Dimbulb. Now rad-hard (*packaging*?) has something to do with cracks or metastability? ...or this thread? You are a piece of work, and still AlwaysWrong.

Look at your legs and consider your mommy's socks. I noticed she's been shopping for undies again.

Just your kind of guy, eh Dimbulb?

You think someone would employ him? Only a government agency would employ someone who was proud of being AlwaysWrong. Maybe the National Weather Service?

See

So it is immersion gold you are recomending. Thanks for clearing up the difference.

At soume point I will have to do some experiments to see what really happens when I try it. I know from someone else who ordered "gold over copper with no nickel", that there is a chance that the vendor will put the nickel in there in spite of my request.

I am also interested in how thin of a FR4 PCB can be bought reasonably. I know that 0.030 is easy to get and that 0.015 can be had. I may be able to leave a step out of the assembly if I can get a very thin PCB. The problem I am trying to solve by making the FR4 thin is that the FR4 is much too good of a thermal conductor and I have an oven in the middle of the assembly.

Oh-oh. I can't have that nickel in there. Does the same apply for the solder getting down to the copper?

I don't know if immersion gold is possible over copper (no nickel). You need to talk to your board supplier. As someone else mentioned, they stripped the gold off their solder terminals. That seems like a lot of work for pcb pads.

It is simple. A mere electrolysis operation.

Just like we can electroplate gold onto a surface, we can also strip gold molecules away from a surface in the same (nearly) manner.

Nah, Homeland Security.

Thanks, Rich

When I was in the USAF, I was an EWS (Electronic Warfare Systems) tech - the planes would go out on missions and come back broken. Something ALWAYS broke - in fact, there was as special "debriefing" department, where the pilots would come after the missions, still in their nifty flying suits, carrying their helmets, and tell the debriefer all of the crap that broke on that mission. Then we'd fill out work orders, etc. Interestingly, this was in Thailand, supporting the VietNam fiasco, and I swear 90% of the pilots that came into debriefing were stoned. But, then I realized, these are just frat boys fresh out of school, in the land of Thai Stick - of _course_ they were stoned! _Everybody_ was stoned! ;-)

Cheers! Rich

Are we talking about "Puff the Magic Dragon" here, or was that before you guys' time? ;-)

Thanks, Rich

This one?

Cheers! Rich

Idiot. What do you think "spooky" was?

IIRC, that wasn't anybody who mattered :>

I don't think I need to do anything like that. I will have to check the gold over copper issue and whether conductive epoxy sticks to gold well enough. Once of the other engineers brought that issue up today.

I may also skip the solder mask on one of the PCBs since it is working more as a mechanical part than as an electrical part. Someone did comment about the solder running along the trace. But I suspect that it isn't going to be an issue because I have huge trace to trace clearances on that part. The one worry I do have is about the boundary where the solder stops.

The little square, metal "lid" you see on top of modern CPUs that appears as if it is soldered on, is really epoxied on with silver filled epoxy. Epotek H20E.

Gold takes to adhesives quite well, so there are many that would work, but the Epotek stuff is pretty damned good.

Sold as two cup sized containers. Part A and B.

Hold the freaking phone, Maybel. You can't go and dress up as Plainclothes Hippie having been in any kind of military service. Next time try "Rich Grise, Hypocrite".

"Smoke pot, not gun barrels"? :rolleyes:

Tim

-- Deep Fryer: A very philosophical monk. Website @